Impaired hyperemic response is predictive of early access failure

The aim of this study was to demonstrate that hyperemic response is a predictor of access failure. We conducted a review of a prospective database of dialysis access patients with preoperative hyperemia studies from June 1998 to August 2002. These consisted of bilateral brachial artery pressures followed by flow velocity measurements of the brachial artery and radial artery at rest and after 3 min of arm ischemia. Measurements were taken by using a cuff placed above the antecubital fossa and inflated to 20 mmHg above systolic pressure. There were no differences recorded in brachial artery pressures for the bilateral studies. Hyperemic response was entered into a stepwise Cox regression to determine its effect on access failure. Access failure was defined as failure to mature or thrombosis. Accesses were placed according to Dialysis Outcome Quality Intiatives (DOQI) guidelines. Kaplan-Meier survival analysis was performed. Log-rank testing was used to compare patency results. Censored end points were death, renal transplant, and access survival to the end of the study period. Fistulas that failed to mature were considered failures at 3 months. Arteries with a <5 cm/sec increase in peak systolic velocity were defined as nonresponders. The 59 arteries used for dialysis access were divided into two groups on the basis of their hyperemic response in cm/sec. The nonresponders were compared with the remainder of accesses performed. Accesses based on arteries with absent or minimal hyperemic response had significantly lower (p < 0.0005) secondary patencies by Kaplan-Meier analysis. Upon further stratification into radial and brachial arteries, the significant difference in secondary patency remained for radial artery--based accesses (p = 0.024) and approached statistical significance for brachial artery--based accesses (p = 0.057). A significant difference was not seen in primary patencies, indicating that accesses based on arteries with an acceptable hyperemic response are more likely to be salvaged by revisions. A nonresponsive radial artery was not a significant predictor of a nonresponsive brachial artery in the same extremity by binary logistic regression (p = 0.111), and a nonresponsive artery was not a significant predictor of nonresponsiveness in the corresponding artery in the contralateral extremity (p = 0.137). Cox regression analysis revealed that the hyperemic response is a significant predictor of failure to mature or thrombosis. Hyperemic testing is a useful means of evaluating adequate arterial inflow for dialysis access. Reduced or absent hyperemic response is an independent predictor of access failure.